Performance and Security Evaluation of Chaotic Maps for Image Scrambling in Real-Time Applications

Image scrambling is an essential technique in secure image transmission, designed to obscure visual data from unauthorized access while preserving the ability to recover the original content. This study investigates the application of five chaotic maps such as Logistic Map, Arnold Cat Map, Lorenz System, Henon Map and Tent Map for image scrambling. Performance is evaluated using key metrics: Normalized Cross-Correlation (NCC), entropy, Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM) and time complexity. Results indicate that the Arnold Cat Map provides the highest entropy (5.8680) and strong decorrelation (NCC = 0.8715), making it ideal for high-security applications. The Logistic Map, with the fastest execution time (0.0039 seconds), is suitable for real-time applications, offering moderate security and good image quality (PSNR = 11.5170, SSIM = 0.0229). The Lorenz System exhibited the best decorrelation (NCC = 0.8657), but its high computational cost limits its use in time-sensitive environments. By comparing the PSNR and SSIM results, this study highlights the balance between visual quality and security, suggesting that chaotic maps can be tailored for specific s ecurity n eeds. The novelty lies in the systematic comparison of these chaotic maps, providing a comprehensive evaluation for optimizing security and performance.